The changes in birefringence in the rigor to relax transition of single tritonextracted rabbit psoas muscle fibers have been investigated with quantitative polarized light techniques. The total birefringence of rest length fibers in rigor was (1.46 -4-0.08) x 10 -s and increased to (1.67 • 0.05) • 10 -s after Mg-ATP relaxation. Pyrophosphate relaxation increased the total birefringence only slightly, whereas subsequent Mg-ATP relaxation elicited the maximum increase in birefringence. Changes in lattice spacing did not account for the total increase in birefringence during relaxation. Moreover, the increase in total birefringence was attributable to increases in intrinsic birefringence as well as form birefringence. No change in birefringence was exhibited upon exposure to a relaxation solution after myosin extraction.Synthetic myosin filaments were prepared and treated with relaxation and rigor solutions. The negatively stained filaments treated with a rigor solution had gross irregular projections at either end, while the filaments treated with a relaxing solution were more spindle shaped. The results are compatible with the view that the subfragment-2 moieties of myosin angle away from the myosin aggregates (light meromyosin) to permit the attachment of the subfragment-I moieties to actin.It is commonly accepted that the force generation mechanism in vertebrate striated muscle involves the angular movements of the subfragment-I (S-I) moieties upon attachment to the actin filaments (16,15,9,27). However, a simple mechanical interaction between the S-I moieties and actin is not possible since the center-to-center spacing between actin and myosin varies at different sarcomere lengths (18). At rest length (ca. 2.70 /~m) in rabbit psoas muscle, the surface-to-surface spacing between actin and myosin is greater than the largest lengths attributed to the S-I moieties of myosin (16,23,25).H. E. Huxley postulated that the subfragment-2 (S-2) moieties of myosin might function as hinges between the light meromyosin (LMM) backbones and the S-I moieties of myosin. Therefore the movements of S-2 moieties away from the long axis of the myosin backbones would permit the attachment of the S-I moieties at the same orienta-